Bayesian network-based high-speed rail traction system fault prediction method

Abstract

The invention discloses a Bayesian network-based high-speed rail traction system fault prediction method. The method comprises the following steps of: 1, establishing a closed-loop bonding graph model: adding a direct torque control module into an original bonding graph model; 2, building a Bayesian network: firstly determining nodes of a Bayesian network structure, determining the structure of the Bayesian network according to the causal relationship between the nodes provided by the closed-loop bonding graph model, and finally opening a closed loop in the Bayesian network; 3, data preprocessing: filtering and denoising the original signal, extracting features, monotonically screening and extracting deep features, and finally, carrying out k-feature extraction on the features; performingmeans clustering; and 4, fault prediction: carrying out supervised learning by using an EM algorithm, obtaining the state of an online measured value by using the data preprocessing method, and carrying out classification by using a Bayesian network to obtain the residual life range of the equipment. Equipment fault prediction can be realized, and the method is more practical and reliable.

Description

technical field [0001] The invention belongs to the technical field of equipment fault prediction, in particular to a Bayesian network-based high-speed rail traction system fault prediction method. Background technique [0002] Fault prediction starts from the current state of use of the equipment, and combines the structural characteristics, parameters, environmental conditions and historical data of the known prediction object to predict, analyze and judge the future faults of the equipment. Fault prediction has great uncertainty, because the object failure mechanism itself is a random process, and the prediction process itself will also produce errors. Due to the large operating range of my country's high-speed rail, the complex and changeable climatic conditions increase the complexity of its working environment, further increase the uncertainty of fault prediction, and bring great challenges to fault prediction. [0003] Bayesian network is currently one of the most ...

AI Technical Summary

Problems solved by technology

At present, the bond graph model of the open-loop system has been built, but the traction system will be disturbed by many environmental factors in the actual operation process, and the open-loop system will become extremely unstable, so the closed-loop control link in the actual system is essential of

Images